Whole brain wiring diagram of oxytocin system in adult mice

Oxytocin (OT) neurons regulate diverse physiological responses via direct connections with different neural circuits. However, the lack of comprehensive input-output wiring diagrams of OT neurons and their quantitative relationship with OT receptor (OTR) expression presents challenges to understanding circuit specific OT functions. Here, we establish a whole-brain distribution and anatomical connectivity map of OT neurons, and their relationship with OTR expression using cell type specific viral tools and high-resolution 3D mapping methods. We utilize a flatmap to describe OT neuronal expression in four hypothalamic domains including under-characterized OT neurons in the tuberal nucleus. OT neurons in the paraventricular hypothalamus (PVH) broadly project to nine functional circuits that control cognition, brain state, and somatic visceral response. In contrast, OT neurons in the supraoptic (SO) and accessory nuclei have limited central projection to a small subset of the nine circuits. Surprisingly, quantitative comparison between OT output and OTR expression showed no significant correlation across the whole brain, suggesting abundant indirect OT signaling in OTR expressing areas. Unlike output, OT neurons in both the PVH and SO receive similar mono-synaptic inputs from a subset of the nine circuits mainly in the thalamic, hypothalamic, and cerebral nuclei areas. Our results suggest that PVH-OT neurons serve as a central modulator to integrate external and internal information via largely reciprocal connection with the nine circuits while the SO-OT neurons act mainly as unidirectional OT hormonal output. In summary, our OT wiring diagram provides anatomical insights about distinct behavioral functions of OT signaling in the brain. Significance Statement Oxytocin (OT) neurons regulate diverse physiological functions from pro-social behavior to pain sensation via central projection in the brain. Thus, understanding detailed anatomical connectivity of OT neurons can provide insight on circuit specific roles of OT signaling in regulating different physiological functions. Here, we utilize high resolution mapping methods to describe the 3D distribution, mono-synaptic input and long-range output of OT neurons, and their relationship with OT receptor (OTR) expression across the entire mouse brain. We found OT connections with nine functional circuits controlling cognition, brain state, and somatic visceral response. Furthermore, we identified a quantitatively unmatched OT-OTR relationship, suggesting broad indirect OT signaling. Together, our comprehensive OT wiring diagram advances our understanding of circuit specific roles of OT neurons.


Abstract 34 35
Oxytocin (OT) neurons regulate diverse physiological responses via direct connections with 36 different neural circuits. However, the lack of comprehensive input-output wiring diagrams of 37 OT neurons and their quantitative relationship with OT receptor (OTR) expression presents 38 challenges to understanding circuit specific OT functions. Here, we establish a whole-brain 39 distribution and anatomical connectivity map of OT neurons, and their relationship with OTR 40 expression using cell type specific viral tools and high-resolution 3D mapping methods. We 41 utilize a flatmap to describe OT neuronal expression in four hypothalamic domains including 42 under-characterized OT neurons in the tuberal nucleus. OT neurons in the paraventricular 43 hypothalamus (PVH) broadly project to nine functional circuits that control cognition, brain 44 state, and somatic visceral response. In contrast, OT neurons in the supraoptic (SO) and 45 accessory nuclei have limited central projection to a small subset of the nine circuits. 46 Surprisingly, quantitative comparison between OT output and OTR expression showed no 47 significant correlation across the whole brain, suggesting abundant indirect OT signaling in OTR 48 expressing areas. Unlike output, OT neurons in both the PVH and SO receive similar mono-49 synaptic inputs from a subset of the nine circuits mainly in the thalamic, hypothalamic, and 50 cerebral nuclei areas. Our results suggest that PVH-OT neurons serve as a central modulator to 51 integrate external and internal information via largely reciprocal connection with the nine circuits 52 while the SO-OT neurons act mainly as unidirectional OT hormonal output. In summary, our OT 53 wiring diagram provides anatomical insights about distinct behavioral functions of OT signaling 54 in the brain. 55 56 57

Significance Statement 58
Oxytocin (OT) neurons regulate diverse physiological functions from pro-social behavior to pain 59 sensation via central projection in the brain. Thus, understanding detailed anatomical 60 connectivity of OT neurons can provide insight on circuit specific roles of OT signaling in 61 regulating different physiological functions. Here, we utilize high resolution mapping methods to 62 describe the 3D distribution, mono-synaptic input and long-range output of OT neurons, and 63 their relationship with OT receptor (OTR) expression across the entire mouse brain. We found 64 OT connections with nine functional circuits controlling cognition, brain state, and somatic

Introduction 70
Oxytocin (OT) is a highly conserved neuropeptide, playing key roles in regulating social 71 behavior and other physiological functions (Althammer et al., 2018; Jurek and Neumann, 2018; 72 Quintana and Guastella, 2020). Impairment in OT signaling has been heavily implicated in many 73 neurodevelopmental disorders including autism (Francis et al., 2014;Rajamani et al., 2018). 74 Altering OT signaling is being pursued as a potential therapy to alleviate social behavioral 75 deficits in many brain disorders (Guastella and Hickie, 2015). However, our limited neural 76 circuit based understanding of OT signaling in the brain hampers the development of targeted 77 therapeutic approaches aimed at altering specific OT functions without affecting other biological 78 pathways. A comprehensive anatomical understanding of OT neurons would enable integrated 79 neural circuit specific studies to decipher the neural substrate of distinct OT functions. 80 The majority of OT producing neurons are located in the paraventricular nucleus of the 81 hypothalamus (PVH) and the supraoptic nucleus (SO) while fewer OT neurons reside in the 82 extended amygdala (Biag et al., 2012;Madrigal and Jurado, 2021). OT neurons receive input 83 from distinct brain regions (e.g., the thalamus) and integrate sensory input with internal 84 information to release OT in a context dependent manner in order to modulate specific 85 downstream circuitry (Grinevich and Neumann, 2020;Tang et al., 2020). The actions of OT are 86 mainly mediated by a single subtype of the OT receptor (OTR) expressed in distinct brain 87 regions as well as peripheral tissues (Gimpl and  Here, we establish a comprehensive wiring diagram of OT neurons in the mouse brain using a 99 high-resolution quantitative brain mapping method in combination with cell type specific 100 transgenic mice and viral tools. All whole brain datasets are registered in the Allen Common 101 Coordinate Framework (CCF) to facilitate data cross-comparison ( were crossed with Ai14 mice, expressing tdTomato following Cre-mediated recombination (Jax: 115 007914, C57Bl/6 J background). 2 months old C57Bl/6 J mice were used for whole brain tissue 116 clearing and immunostaining. Mice received food and water ad libitum and were housed under 117 constant temperature and light conditions (12 hrs light and 12 hrs dark cycle). 118

Experimental design and statistical analyses 119
For OT neuron distribution mapping (Fig.1), we used 3 males, 3 females (virgin), and 2 females 120 (lactating) of 2 -4 month old OT-Cre;Ai14 mice with STPT imaging. We also used 4 males, 3 121 females (virgin) of 2 month old C57bl/6 mice for tissue clearing and LSFM imaging based 122 quantification (Fig.1). Since we did not observe significant difference in OT neuronal number, 123 we combined data from both sexes to generate representative cell counting (Table 1). For 124 anterograde projectome mapping in 2 -4 month old OT-Cre (Fig.2), we used 2 males, 3 females 125 (virgin), and 3 females (lactating) with 500nl of AAV injection, and 3 males, 3 females (virgin), 126 and 1 female (lactating) with 50-150nl of AAV injection for the PVH targeting. Moreover, we 127 used 5 males and 5 females (virgin) for the SO, 3 males and 2 females (virgin) for the TU, 1 128 male and 1 female (virgin) for the AN. For oxytocin receptor expression mapping using OTR-129 Venus mice (Fig.3), we used 6 males and 8 females (virgin). For rabies input mapping (Fig.4), 130 we used 3 males and 3 females (virgin) for the PVH, and 3 males and 1 female (virgin) for the 131 SO. 132 To determine the correlation between OT area normalized projection and OTR density (Fig. 3C), 133 we first tested for the normality of the data using the D' et al., 2020). Briefly, the brain was embedded in 4% oxidized agarose and cross-linked with 185 0.2% sodium borohydride solution. The brain was imaged as 12 x 16 x 280 tiles with 1 x 1 µm 2 186 x,y pixel resolution in every 50 µm z-section. We used 910 nm wavelength for two-photon 187 excitation to excite both green (e.g., eGFP) and red signals (e.g., tdTomato). Signals were 188 separated with 560 nm dichroic mirror and two band path filters (607/70-25 for red and 520/35-189 25 for green). Imaging tiles in each channel were stitched with custom-built software (Kim et al.,190 2017; Newmaster et al., 2020). 191 For quantitative projection data analysis, we used our previously published pipeline (Jeong et al.,192 2016). Briefly, both signal and background channels were z-normalized. Then, the background 193 channel images were subtracted from the signal channel images to increase signal-to-noise ratio. 194 Then, projection signals were converted to a binary map by applying an optimized threshold (8x 195 standard deviation) to detect signals while minimizing noise from background autofluorescence. 196 Then, binarized signals in each pixel were counted in 20 x 20 (x,y) pixel unit (voxel) and the 197 value was assigned the corresponding voxel across the brain, which is defined as "projection 198 strength". Thus, range of the projection strength in a given voxel is between 0 and 400. 199 Projection strength of each area is calculated by summing up all projection strength within an 200 anatomically defined area. Autofluorescence of brains was used to register each brain to the 201 Allen CCF using Elastix (Klein et al., 2010), then, the projection signals were transformed to the 202 reference brain. Then, we used maximum projection of registered long-range output datasets 203 from each area to create a representative projection data for further quantitative analysis (Movie 204 S2). "Area normalized projection" represents normalized occupancy of projection signals in the 205 ROI by dividing the projection strength with a total number of voxels in each ROI. For example, 206 if total voxel count for one ROI was 20,000 and our projection strength showed 2,000 in the 207 ROI, it will be (2,000/20,000)*100 = 10%. Regions with a projection strength greater than 1% is 208 designated as dense, between 1 and 0.5 as intermediate, between 0.5 and 0.1 as sparse, and less 209 than 0.1 as very sparse (Fig. 2E). calculated by dividing cell number with ROI area. 2D counting numbers were also converted 213 into 3D counting using our previously calculated 3D conversion factor (1.4 for tdTomato) (Kim 214 et al., 2017). To measure the volume of anatomical ROI, the reference Allen CCF was reverse 215 registered onto individual brains using the Elastix. "Cell density (counts/mm 3 )" was calculated 216 by dividing detected cell numbers in 3D with the anatomical ROI volume. The cell counting 217 analysis was applied to OT-Cre;Ai14 and OTR-Venus cell distribution and inputs to the OT 218 neurons. We used an average of individual datasets to create representative OT (OT-Cre;Ai14, 219 Movie S1) and OTR (OTR-Venus, Movie S3) distribution and maximum projections to create 220 mono synaptic input for OT neurons (rabies). For rabies input degree (Fig. 4D, Movie S4), 221 Regions more than 100 cells are designated as dense, between 100 and 10 as intermediate, and 222 less than 10 as sparse. 223 To compare relative abundance between OT output and OTR expression in Fig. 3C., relative cell 224 density or output data in each region was calculated by dividing each data by summed density or 225 output data from all areas (excluding viral injection sites), respectively. Then, log10 (relative OT 226 output/relative OTR) was used to examine the quantitative relationship between the two signals. 227

2D hypothalamic and PVH Flatmap 228
To generate the hypothalamic flatmap, we adapted the previously used method (Kim et al., 2017) 229 and applied it to the hypothalamic region. First, we created a binary image in the hypothalamic 230 area based on the oxytocin expression. Second, a zero line was placed to generate evenly spaced 231 bins along the dorsal to the ventral direction of the PVH and laterally extended to include TU 232 and MEA at different coronal plains. To capture signals on the flatmap, bins were registered into 233 the reference brain and the cell number in each bin was quantified as described before in the 234 STPT data analyses section. Lastly, the mean number of the OT neurons in 8 OT-Cre;Ai14 235 brains were plotted in each flatmap using a custom-built matlab code. For the PVH flatmap, we 236 followed the same procedure to generate a hypothalamic flatmap except for bin generation. 237 Instead of delineating bins in a binary image, we assigned bin numbers in the PVH subregion of 238 Franklin-Paxinos atlas (Paxinos and Franklin, 2008) in the dorsal to the ventral direction. 239 240 Whole brain clearing and immunostaining, light sheet microscopy, and cell counting 241 C57Bl/6 J mice (4 males and 3 females at P56) were transcardially perfused with 0.9% saline 242 followed by 4% PFA in 0.1M phosphate buffer (PB, pH 7.4). The decapitated heads were 243 postfixed in 4% PFA overnight at 4°C and brains were dissected out the following day. All the 244 following steps were performed on an orbital shaker unless otherwise specified. Dissected brains 245 were delipidated in SBiP buffer (0.2mM Na2HPO4, 0.08% SDS (Sodium Dodecyl Sulfate), 246 0.16% 2-methyl 2-butanol, 0.08% 2-propanol). Delipidation was performed with 3-4 washes 247 (10ml per wash) in SBiP for 24 hrs followed by one 10ml wash with SBiP for the next 4 days. 248 Samples were then moved to B1n buffer (0.1%v/v Tritox-X-100, 1% wt/v glycine, 0.001N 249 NaOH, 0.008% wt/v sodium azide) for 1 day (10ml) and then shifted to 37 0 C incubation for 3 250 hrs. Once delipidation was completed, the samples were washed in PTwH (Tween 20-2ml, 251 10mg/ml Heparin-1ml and sodium azide-2g, made to 1L with 0.1M phosphate buffered saline) 252 3-5 times at 37 0 C for 24 hrs. The samples were then incubated in antibody solution (5% DMSO 253 and 3% Donkey serum in PTwH-4ml per sample) containing primary antibodies for OT 254 (ImmunoStar Cat# 20068, RRID:AB_572258, 1:500) and Vasopressin (Peninsula Laboratories 255 Cat#T-5048, RRID:AB_2313978, 1:1000) for 10 days at 37 0 C. Next, PTwH washes were 256 performed 4-5 times for 24 hrs at 37 0 C, followed by secondary antibody incubation. Secondary completed, the samples were moved to room temperature (RT) and further processed for tissue 263 clearing. All the following steps were performed in a fume hood in glass containers and the 264 containers were filled completely. Samples were dehydrated in the following series of methanol laboratories, H-1500-10). For microscopic imaging, a BZ-X700 fluorescence microscope 290 (Keyence) and a confocal microscope (Zeiss 510) were used. A low magnification objective lens 291 (4x) was used to image with a large enough view to define brain anterior-posterior location from 292 bregma and higher magnification objective lenses (10x ~ 40x) were used to image sections 293 depending on the cell density. Images were delineated manually based on the Franklin-Paxinos 294 atlas and fluorescently tagged cells were manually quantified using the cell counter plug-in in 295 FIJI (ImageJ, NIH). 296 297

Software Accessibility 298
All custom-built codes and flatmaps used in the current study will be freely available upon 299 request and can be used without any restriction. 300 301

Data Sharing Plan 302
Data files for the anterograde projectome, rabies based monosynaptic input, and OTR expression 303 data registered on the Allen CCF are included as supplementary data. 304 High-resolution serial two-photon tomography images will be deposited in BrainImageLibrary 305 (https://www.brainimagelibrary.org/) and web visualization link will be added upon publication. 306 307

Results 308
Quantitative density mapping of oxytocin neurons reveals four clusters in the adult mouse 309 brain 310 We first aim to determine quantitative brain-wide OT distribution in complex 3D structures. To 311 examine the anatomical distribution of OT neurons across the whole brain, we used OT knock-in 312 mice with Cre recombinase (Ot-Cre) crossed with Ai14 reporter mice (OT-Cre;Ai14-313 heterozygotes) (Choe et al., 2015). We imaged the entire mouse brain at cellular resolution using 314 serial two-photon tomography (STPT) and performed quantitative mapping using previously 315 established computational methods (n=8 brains, Fig. 1A pattern of OT neurons, we created a flatmap (Fig. 1C). Evenly spaced bins provide a flattened 2D 320 spatial unit to quantify and to display signals from the 3D brain. The flatmap was delineated with 321 Allen Common Coordinate Framework (Allen CCF) and Franklin-Paxinos atlas based anatomical 322 labels ( 4. the TU area (Fig. 1E) (Knobloch and Grinevich, 2014). Notably, the largely overlooked TU area 328 contains almost as high density of OT neurons as the PVH area (Fig. 1E). 329 To distinguish neurons actively expressing OT in adults from developmentally labeled cells, 330 we performed immunohistochemistry using an OT antibody in OT-Cre;Ai14 mice. We confirmed 331 that almost all OT immuno positive neurons (97%, 1733 out of 1790 cells, n=4 animals) were 332 labeled by tdTomato from OT-Cre;Ai14 mice ( Fig. S1). In contrast, 76% of tdTomato labeled cells 333 were OT immuno positive (1733 out of 2277 cells) in the PVH. Smaller portions of tdTomato cells 334 in the SO (44%, 654 out of 1508 cells) and the MEA (8%, 31 out of 375 cells) retain active OT 335 expression (Fig. S1). This result suggested that OT neurons undergo OT expression changes during 336 neurodevelopmental processes (Madrigal and Jurado, 2021). 337 To cross validate active expression of OT in the adult brain, we performed tissue clearing 338 followed by 3D immunolabeling with OT and vasopressin antibodies in 8 weeks old C57bl/6 mice 339 (n= 7 brains; Fig. 1F-I) (Renier et al., 2016). We developed 3D counting and image-registration 340 methods to achieve similar unbiased brain-wide cell counting as done with STPT imaging (see 341 Methods for more details). We observed similar OT staining distribution and overall slightly 342 higher counting compared to our transgenic based mapping results (Table 1). For example, the 343 estimated number of OT neurons in the PVH with the immunostaining was 1,095 cells out of total 344 3149 cells (~35%), which is higher than our transgenic based estimate mostly likely due to 345 sensitive labeling based on antibody detection. Importantly, we confirmed the robust OT 346 expression in the TU area that was not colocalized with vasopressin staining (Fig. 1I). 347 348 Quantitative whole brain projection mapping of OT neurons reveals broad long-range 349 projections in nine functional circuits 350 Next, we aim to establish a comprehensive anterograde projection map from OT neurons in the 351 four identified areas and examine whether OT projections target specific functional circuits related 352 to distinct behavior control. 353 Since OT can be released via axons, dendrites, and even neuronal processes (Jurek and 354 Neumann, 2018), we injected a Cre-dependent adeno associated virus 2 (AAV2-CAG-Flex-EGFP) 355 in the four areas of Ot-Cre knock-in mice with slightly varying injection sites to cover target areas 356 (N = 15 animals for the PVH, 10 for the SO, 2 for the AN, and 5 for the TU) ( Fig. 2A-B). We 357 included male, virgin female, and lactating female mice in our study (see method for more detail), 358 and observed no significant difference between sex or lactating state. Thus, we merged all data 359 from the same anatomical areas. Long-range projection signals from individual injections were 360 then registered onto the Allen CCF and maximum projection data in all samples from each 361 anatomical area were used to represent efferent output for the four areas ( Fig. 2C; Movie S2). We 362 found abundant projections from OT neurons in the PVH to the midbrain and hindbrain areas while 363 relatively sparse projection to the diencephalon and telencephalon areas (Fig. 2D) 364 We then examined whether OT neurons in the four anatomical areas show any distinct 365 projection pattern. Overall, the PVH neurons showed the broadest projection pattern followed by 366 the SO and the AN, which project to a small subset of PVH-OT efferent areas ( Fig. 2D-E; Movie 367 S2). The TU-OT neurons did not show any long-range projections. We ask whether OT neurons 368 project to distinct neural circuits related to specific function. Based on known functions of each 369 anatomical region, we found that PVH-OT neurons project to three functional modules that control 370 the internal state, somatic visceral, and cognitive response. Each module contains three circuits. 371 The internal state module contains attention, threat/alert/defense, and sleep/awake circuits (Fig.  372  2E). The somatic visceral module includes pain, sensory motor, and body physiology/metabolism 373 circuits (Fig. 2E). Lastly, the cognitive control module has learning/memory, reward/value 374 assessment, and reproduction circuits (Fig. 2E). Each circuit is composed of multiple brain regions 375 from the hindbrain, midbrain, thalamus, hypothalamus, cerebral nuclei, and cerebral cortex that 376 process low-to-high order information. For instance, many basal ganglia circuit components 377 including the caudate putamen (CP), globus pallidus (GP), subthalamic nucleus (STN), and 378 substantia nigra (SN) receive OT projection to modulate motor function (see sensory motor 379 regulation in Fig. 2E). PVH-OT neurons project to these areas at varying degrees. Dense projection 380 occurs largely onto the hindbrain (e.g., the dorsal motor nucleus of the vagus nerve; DMX, the 381 parabrachial nucleus; PB), the midbrain (the substantia nigra compacta; SNc), and the 382 hypothalamus (the medial preoptic nucleus; MPN) to directly modulate motor output and sensory 383 input (Fig. 2E). In contrast, the cerebrum (the cerebral nuclei and cerebral cortex) that works as 384 high cognitive controller receives more sparse projection (Fig. 2E). The SO and AN project to a 385 small subset of PVH-OT target areas (Fig. 2E). This data suggests that OT neurons in these two 386 areas can further modulate a subset of the nine functional circuits, albeit less powerfully. 387 Together, our comprehensive projectome analysis uncovers anatomical substrates to 388 explain pleiotropic effect of OT neurons regulating diverse behavioral outcomes.

390
Oxytocin receptor expression showed quantitative mismatch with oxytocin neuronal output 391 Next, we ask whether expression of a single subtype of the OT receptor (OTR) is quantitively 392 correlated with OT projection target areas to mediate circuit specific OT function. Although most 393 OT projecting areas are known to contain OTR expression (Grinevich et al., 2016), the quantitative 394 relationship between OT projection and OTR expression across the whole brain is currently 395 lacking.

396
To understand OT-OTR correlation, max OT projectome data from both the PVH and SO 397 were compared to OTR expression in adult mice using a previously validated mouse line, OTR-398 Venus (Newmaster et al., 2020). A cohort of adult OTR-Venus mice brains were imaged using 399 STPT and mapped OTR expression in the whole adult brain. These mapped OTR positive neurons 400 (magenta in Fig. 3A) were registered onto the same reference brain along with OT-projections 401 (green in Fig. 3A). Overall, the OTR showed high expression in the cortical area with minimal OT 402 projection, while many midbrain and hindbrain regions have strong OT with little OTR expression 403 (Fig. 3A-B; Movie S3). When we examined whether relative projection of OT neurons is 404 correlated with relative OTR density across the entire brain, we found no significant correlation 405 across the whole brain and major brain areas, except for the thalamus and the medulla (Fig. 3C). 406 Overall, our results highlight lack of quantitative and spatial correlation between OT projections 407 and OTR expression in the mouse brain. 408 Then, how do OTR rich areas (e.g., the isocortex) receive OT signaling without direct OT 409 projection? A previous study suggested that many OTR expressing neurons may receive OT signal 410 non-synaptically via cerebral spinal fluid (CSF) (Zheng et al., 2014). Hence, we examined whether 411 OT projection fibers make physical contact with ventricles. Indeed, we frequently found OT fibers 412 with thick varicosities at the lateral, 3 rd , and 4 th ventricle surface (Fig. 3D). This further suggests 413 that OT signaling may transmit to the brain via the CSF route in addition to direct transmission in 414 target areas.

416
OT neurons mainly receive monosynaptic inputs from the thalamus, hypothalamus, and 417 cerebral nuclei. 418 Since OT neurons are known to integrate external stimuli and internal state, we ask whether OT 419 neurons in the PVH and the SO receive mono-synaptic input from sensory and integrative 420 information processing brain areas. 421 To map brain-wide mono-synaptic inputs in a cell type specific manner, conditional 422 retrograde pseudorabies viruses were injected into the PVH and the SO of the Ot-Cre knock-in 423 mice separately (Fig. 4A) (Wickersham et al., 2007). We confirmed the specificity of labeling by 424 performing co-immunolabeling TVA positive neurons with OT and AVP. None of the TVA 425 infected neurons were AVP positive and were largely OT positive (Fig. S2). We also confirmed 426 no leakiness of TVA labeling by injecting TVA in the PVH of adult C57 mice which did not result 427 in any infection (Fig. S2). To confirm G protein dependency for monosynaptic tracing, TVA 428 without G and rabies viruses were injected to the PVH and SO separately and the brains were 429 imaged in STPT (N=1 animal each for the PVH and SO). All the neurons observed were confined 430 to the injection site and both samples were devoid of any long-range input cells. Lastly, we 431 performed another rabies tracing experiment with optimized G and split TVA that are known for 432 improved Cre specificity and tracing . We found near identical results with this 433 alternative virus approach (Fig. S3) (N=2 animals, each for the PVH and SO). Once we confirmed 434 the validity of our input tracing methods, we used our mapping method to quantify input neurons 435 throughout the whole brain (Kim et al., 2017). To acquire overall inputs to each anatomical area, 436 input signals from multiple independent injections targeting a specific brain region were registered 437 onto the Allen CCF and the max projection of input signals from each anatomical area (N = 6 438 animals for the PVH and 4 for the SO) were overlaid onto the reference brain (pseudo-colored 439 green for the PVH and magenta for the SO in Fig. 4B-C; Movie S4). 440 Overall, OT neurons from the PVH mainly receive inputs from the thalamus, hypothalamus, 441 and cerebral nuclei (Fig. 4C). All brain regions providing inputs to the OT neurons also receive 442 output from the OT neurons except the triangular nucleus of septum (TRS), creating reciprocal 443 connections with afferent areas (Fig. 2E and 4D). Noticeably, OT neurons received little input 444 from hindbrain despite strong output to the same area, suggesting that OT neurons provide largely 445 unilateral output to the hindbrain (Fig. 2E and 4D). Moreover, the cerebral cortex provides little 446 to no input to the OT neurons, further supporting very weak direct interaction between cerebral 447 cortical areas and OT neurons ( Fig. 4D; Movie S4).

448
SO-OT neurons receive overall similar input compared to the PVH-OT neurons (Fig. 4D). 449 The broad afferent pattern is in sharp contrast to the very sparse efferent projection of SO-OT 450 neurons to the brain (Fig. 2E and 4D). When monosynaptic input from the PVH-and SO-OT 451 neurons are compared, SO-OT neurons show input from relatively more lateral parts of the brain 452 (Fig. 4C, arrows). 453 Collectively, we conclude that OT neurons in the PVH and the SO receive similar input 454 from a subset of brain areas that receive majority of input from hypothalamic areas followed by 455 cerebral nuclei and thalamic areas. (Fig. 4D).

457
Input-output wiring diagrams of PVH-and SO-OT neurons provide overall neural circuit 458 control patterns. 459 Based on our long-range output and mono-synaptic input data, we constructed input-output circuit 460 diagrams of OT neurons in the PVH and the SO while annotating each brain area based on their 461 functional categories (Fig. 5). PVH-OT neurons project broadly to all nine identified functional 462 circuits throughout the brain, indicating that OT neurons can modulate information processing at 463 different level of circuits with overall stronger influence in the mid-and hindbrain circuits (Fig.  464  5). In contrast, mid-level circuits including the diencephalon (the thalamus, hypothalamus), the 465 midbrain, and the cerebral nuclei, provide major input to inform action of OT neurons, providing 466 anatomical substrate to perform an integrative role (Fig. 5). SO-OT neurons receive similar mid-467 level circuit input compared to the PVH-OT neurons while showing limited central projection to 468 the midbrain and pons (Fig. 5). This suggests that SO-OT neurons mainly serve as peripheral 469 hormonal output. 470 471

Discussion 472
The wiring diagram of the brain is a structural foundation to decipher neural circuits underlying 473 brain function. Here, we present a comprehensive anatomical connectivity map of the 474 hypothalamic OT neurons and their relationship with postsynaptic OTR expression in the whole 475 mouse brain. A quantitative mismatch exists between OT projection and OTR distribution pointing 476 towards abundant non-synaptic OT signaling within the brain. We also identify nine functional 477 circuits with reciprocal or unidirectional connection with OT neurons that serve as anatomical 478 entities to exert varied behavioral control. 479 480 OT neurons are mostly located in hypothalamic nuclei with a complex 3D shape (Biag et al., 2012;481 Madrigal and Jurado, 2021). To examine OT expression intuitively and quantitatively, we devised 482 a 2D flatmap for OT containing hypothalamic regions from an Allen CCF based reference brain 483 while incorporating anatomical labels from the Allen Institute and Franklin-Paxinos provides an alternative coordinate system to understand anatomical connectivity. In addition to 487 well-described OT neurons in the PVH, SO, and AN, we described another major population in 488 the TU area in the hypothalamus (Jirikowski, 2019). Our 3D immunolabeling independently 489 validated the existence of this extra population. Our anterograde tracing showed that these neurons 490 have almost no central projection, suggesting their contribution to brain information processing is 491 limited. Future studies including ablation studies will help to elucidate the functional significance 492 of this overlooked OT population.

494
OT signaling is known to modulate many distinct brain functions such as anxiolytic effect, social known functions, we identified nine functional circuits where OT processes interact to modulate 498 distinct behavioral circuits. Each circuit consists of a set of brain regions processing different 499 behavioral aspects. Thus, our circuit map can help to understand neural entities of OT that 500 modulate different behavioral aspects. Overall, OT circuits provide broad projections to modulate 501 external and internal information throughout the entire brain circuit. For example, we found that 502 OT neurons project to sensory-motor and pain circuits from the hindbrain and midbrain to cerebral 503 cortex and cerebral nuclei. A recent single cell reconstruction study demonstrated that even a single 504 magnocellular OT neuron can make multiple collateral projections to extra-hypothalamic areas to 505 coordinate neuromodulation across functionally related brain circuits (Zhang et al., 2020). These 506 provide anatomical evidence that OT neurons can finely modulate sensory motor processing 507 throughout different circuit levels. Notably, OT neurons project to other neuromodulatory areas 508 such as the locus coeruleus for norepinephrine (alert), the substantia nigra (movement) and the 509 ventral tegmental areas for dopamine (reward), and raphe nuclei for serotonin (emotion), thus Social behavior is a complex behavior, requiring coordinated interplay between the sensory system 514 and integrative circuits to generate socially appropriate motor outputs. Extensive connections of 515 OT neurons to somatic visceral, cognitive, and state control modules can help to fine-tune activity 516 of different circuit components to generate enhanced response to socially salient stimuli. 517

518
OT gets released through axonal and dendritic projections based on the inputs that OT neurons 519 receive. The presence of large dense core vesicles containing OT at the non-active zones of pre-520 synapses (Theodosis, 1985;Griffin et al., 2010), absence of evidence for OTR in the postsynaptic 521 membranes, and extremely delayed electrophysiological OT response (milliseconds to seconds) 522 (Knobloch et al., 2012;Knobloch and Grinevich, 2014) collectively support non-synaptic axo-523 dendritic release of OT. Hence, our projection maps with entire process labeling provide possible 524 release sites of OT throughout the whole brain. We also compared OT total projections (combined 525 data from the PVH-and SO-OT neurons) to OTR expression in the central brain. Our OT mono-synaptic input maps showed that the majority of inputs are from the cerebral nuclei, 544 thalamus, hypothalamus, and midbrain with little input from the hindbrain. Particularly, almost all 545 afferent brain regions to PVH-OT neurons also receive efferent projections, suggesting strong 546 reciprocal control of target regions by PVH-OT neurons except the hindbrain for unidirectional 547 output. Abundant bidirectional connections with nine functional circuits suggest that PVH-OT 548 neurons can be an allostatic tool to interactively orchestrate and facilitate social and non-social 549 information processing based on external stimuli and internal state (Quintana and Guastella, 2020). 550 In contrast, despite having similar afferent areas to SO-OT neurons, their limited central projection 551 suggest that SO-OT neurons serve largely as unidirectional hormonal output to the periphery rather 552 than reciprocal circuit modulator. 553

554
In summary, our study provides an anatomical foundation to understand diverse functions based 555 on OT neurons in the brain. We deposit all high-resolution imaging data in publicly accessible 556 databases and our website to facilitate data mining. We envision that this OT wiring diagram with 557 quantitative expression data will guide future studies to understand circuit-based mechanisms of 558 OT function and its changes in socially relevant behaviors as well as brain disorders such as autism.  circuits that provide mono-synaptic input to OT neurons in the two anatomical areas. Note 773 overall similar input patterns for both areas. The full name of abbreviations can be found in 774 Transgenic animal counting is from serial two-photon tomography imaging of OT-Cre;Ai14 792 mice (n=8) and 3D immunolabeling counting is from light sheet fluorescence microscopy 793 imaging of C57 after tissue clearing and 3D oxytocin antibody staining (n=7